Mid-Level Co ntroller
2.6 Conclusions
Power transformers are specifi ed to withstand the me-chanical forces arising from both shipping and subsequent in-service short circuits across the terminals. Th e most severe service forces arise from close in system faults, faults in a load tap changer and, for a generator transformer, energiz-ing out of synchronization. Short circuit forces produce axial and radial forces and these can lead to radial buckling or axial deformation (twisting, displacement of clamps or supports). Transport damage can occur if the clamping and restraints are inadequate, leading to core and winding movement. With a core form design, the principal forces are in the radial direction, while a shell form design is in an axial direction. Th is diff erence is likely to infl uence the types of damage found.
The technology assisting transformer designers has improved over recent years, but it is rare for the designs to be evaluated other than by subsequent service life. Once a unit has been damaged, even if only slightly, the ability to withstand further short circuits is reduced. Th e requirement is to have eff ective methods of identifying damage. One ap-proach is to rely upon an internal Visual inspection, but it is invariably too diffi cult to draw eff ective conclusions. Th e oil has to be drained and confi ned entry rules apply. Since so little of the winding is visible, often little is seen other than displaced support blocks. Consequently, the reliance must be on condition assessment methods. However, since the consequences of an incorrect diagnosis are so great, a mandate is to have a range of complementary and eff ective diagnostic techniques available for fi eld use.
Th e requirement is to identify damage of the following types:
• Short circuit turns
• Open circuits
• Core ground problems
• Core movement
• Axial or radial deformation
Mechanical Condition
Condition assessment evaluation of transformers would be carried out on the following occasions:
• During an investigation, after a fault or protection trip. Th e purpose would be to determine the nature and extent of any damage.
• During a condition assessment. Th is test may done as part of a general assessment, or the unit may be known to have seen short circuits over time, and ap-parently successfully withstood them. In this latter case the test would be to identify possible damage and used to indicate the capability to withstand fur-ther short circuits.
• Before and after a relocation. Comparisons of test data made before and after a relocation, which should indicate any mechanical movement.
• By manufacturers as a quality check of the manufac-turing process, by comparing the response of units made to the same design.
• Testing is also carried out on new and refurbished units to obtain fi ngerprint values for references. Also, test results on sister units (similar design) can be used as references.
Consequences of Diagnostic Testing
The result of such testing may have a number of implications:
• If the test indicates damage or malfunction, and the test has been performed after operation of a protec-tion relay - the unit is likely to need a major repair or scrapping. Further, confi rmatory evidence may be necessary (e.g. additional testing specifi c to the type of fault indicated).
• If there is evidence of some damage or deformation, but there are no other signs of malfunction - the unit may be returned to service. Engineering judgment is required to review the risk of failure at the next short circuit, the likelihood of such an event, and the sys-tem risk exposure. Th e results would be stored and used as a benchmark indicative of worsening of the damage.
• Where there is no evidence of damage or deforma-tion, and there is no other evidence (or expectation) of a malfunction - the unit is validated for service and the results archived for future use.
Test Program
Th e following tools would be used:
• Insulation Analyzer - to measure capacitance and power factor, exciting current and turns ratio.
• Leakage Reactance Interface - to measure short circuit impedance.
• Sweep Frequency Response Analyzer - to measure the transfer function.
• Winding and insulation resistance.
• Other test data relating to the period prior to de-energization could be relevant - such as dissolved gases and furans from an oil sample, Infrared and RIV scanning (PD).
Th e transformer would be de-energized and all high volt-age connections removed. Th e circuit and the transformer should be made safe for testing, according to standard com-pany procedures. Ideally the transformer will have normal service oil in the tank. For the test program it is necessary to remove any temporary bushing ground connections. Th e leakage reactance and SFRA tests also require removing grounds from neutral bushings. A transformer with an off load tap changer would be tested in its normal operating position. A unit with a load tap changer would normally be tested in an off -neutral position and preferably throughout its full range.
Assessment
While the objective is to assess the mechanical condition, the test data would be used to provide a more general assess-ment - of the insulation condition for example. Specifi cally, however, the following methods would be applicable to the mechanical assessment:
• Winding Capacitance
Th e Doble M4000 Automated Insulation Ana-lyzer can be used to measure winding movement, and is probably the most commonly used of all the methods. Th e technique is capable of detect-ing gross winddetect-ing movement. In addition, since the capacitance of a low voltage winding is mea-sured to ground, it is sensitive to disruption of the core ground connection, and will detect gross core movement. Th e sensitivity can be enhanced, where it is possible, to make separate measurements on each phase and so use inter-phase comparisons.
With autotransformers, it is not possible to mea-sure inter-winding capacitances between high and low voltage windings.
• Exciting (or Magnetizing) Currents
Th e Automated Insulation Analyzer can be used to measure exciting currents and watts loss. Th is can be one of the simplest methods to detect shorted turns, following a short circuit. It can also detect open and short circuits elsewhere - in the LTC, core and core ground. It is a comparative method with most of the supporting documents appear-ing in the 1970’s (9) where evidence was presented that it can identify a range of core related features - shorted laminations or fundamental changes in the iron characteristics.
• Leakage Reactance/ Short Circuit Impedance Standards for short circuit testing of
transform-ers usually specify this measurement. It involves a simple interpretation of a change in one value to another and is very suitable for a contractual use in a highly controlled environment. During factory acceptance the impedance is measured with three-phase excitation and high currents. Field test are usually single phase and at a low current. To relate the measurements it is necessary to undertake the procedure according to the Doble method and the M4110 Leakage Reactance Interface uses this ap-proach (10). Experience indicates that an accuracy of around 0.2% is needed to detect a 0.5% change over nameplate values. Th e success of the method relies upon the availability and reliability of factory data. In some cases a phase-by-phase comparison may assist in the analysis.
• Sweep Frequency Response Analysis
Th ere is a direct relationship between the geomet-ric confi guration of the winding and core and the series and parallel impedance network of induc-tance, capacitance and resistance. Th is network can be identifi ed by its frequency-dependent transfer function. Frequency Response Analysis testing by the sweep frequency method (SFRA) uses network analysis tools to determine the transfer function. Changes in the geometric confi guration alter the impedance network, and in turn alter the transfer function. Th is enables a wide range of fail-ure modes to be identifi ed.
Doble uses the protocols developed by the Eu-roDoble Client Group. From this base, Doble has subsequently developed an instrument to match the requirements, the M5100 SFRA. Th e SFRA method is also comparative between phases and against previous results. Th ere is also some com-monality between units of the same design.
Sweep Frequency Response Analysis
A general impedance diagram for a transformer is shown in Figure 1.
Figure 1 — Transformer Impedance Model
Th e transfer function approach is to consider a trans-former as though it was a simple inductance, capacitance and resistance (L-C-R) equivalent circuit and determine its frequency admittance response.
Th e basic measurement formula for the transfer func-tion is:
Attenuation = 20*log (Vout/Vin) for all frequencies.
At low frequencies the impedance ladder is represented by the series inductance and winding resistance. At medium frequencies the capacitance to ground is relevant, and at higher frequencies the relevant impedances are the series and ground capacitances.
Much of the past work has been done using a laboratory instrument – a super heterodyne network analyzer used over a 10Hz to 10MHz range of frequencies. Th e Doble M5100 SFRA Instrument has been developed to meet the applica-tion requirement however; it is enhanced by the simplicity of a single function, automated control, data storage, fi eld ruggedness and noise immunity. All of the features required for substation test instrumentation.
Figure 2 shows a circuit diagram of the M5100 SFRA Instrument. It has the following characteristics:
Figure 2 — M5100 SFRA Circuit Diagram
Measurement).
Th e transformer test involves applying a test signal to one terminal of the transformer under test and measur-ing this applied signal at the same terminal, and also the signal appearing at a second terminal, as shown in Fig 2.
Signals are applied and measured with respect to ground.
Th e amplitudes and phases of the two signals, S Measure-ment and R MeasureMeasure-ment, are measured to determine the relative amplitude and phase shift changes between them.
Th e basic measurement is of the attenuation and phase shift of a signal after having passed through the winding from the input to the output terminal. Th e test can also include voltage transfers between windings i.e. applying a signal to one winding of a transformer and measuring the response at another winding to determine the amplitude change and phase shift of the signal having been transferred along a winding, or from one winding to the other.
Early attempts to gain repeatability, particularly using impulse methods, were not successful. Th e success of the SFRA method is the result of a signifi cant eff ort in develop-ing a common protocol by EuroDoble Clients.
While the application is now fairly straightforward, interpretation requires experience to diagnose the type of fault. Shown in Figure 3 is a typical set of results for an autotransformer in good condition. For most transformers there is a large attenuation at a specifi c low frequency, usually between 400 – 1500Hz. Below this frequency, the impedance is dominated by the series inductance and measurement resistance of 50 Ohms. Since the impedance is controlled by the core magnetization, this is where core effects are seen and there is some equivalence with an ex-citation current measurement. The center phase response is slightly different in this area of frequency, due to the different fl ux paths through the core. In addition, the center phase has a single null, shown at 600 Hz and the two outer phases overlap with a double resonance around the same frequency. At this frequency, there is a phase change of 180 degrees and the impedance changes from being inductive to capacitive domination. At higher frequencies, in kilo and megahertz ranges, eddy currents shield the magnetic circuit and local leakage fl uxes determine the winding inductances.
The response is more dependent upon changes in the wind-ing, and the diagnostics should compare with the leakage reactance measurements.
Figure 3 — A Set of Normal Test Results from an Autotransformer
Figures 4 and 5 show the results from damaged units.
Experience shows that diff erences in the lower frequency ranges relate to core changes, or shorted/open circuits.
Medium frequencies show winding shifts, while more localized winding movement is seen at the higher frequen-cies. In the result shown in Fig.4 there are two phases that overlay with a minimum at 400 Hz and again at 2200 Hz however, the third (red trace) does not follow the same pattern, as it should. It’s minimum has shifted indicating a problem. Figure 5 also has identical resonances on only two of the phases. Experience indicates that changes of this type, at these frequencies are associated with winding deformation.
Figure 4 — Test Results Indicating Shorted Windings
Figure 5 — A Transformer With Axial Deformation
Conclusions
A power transformer is one of the most critical items in a power system. It also has a very high capital value. In order to achieve the full benefi t of this asset, it is important to have the most eff ective means of identifying any deteriora-tion or malfuncdeteriora-tion. Visual inspecdeteriora-tions are not as eff ective as on other types of apparatus, such as circuit breakers, yet expensive decisions often have to be made relating to the future serviceability. Th is can only be achieved through the application of a broad range of complementary assessment tools. Within this context, Sweep Frequency Response Analysis with instruments such as the Doble M5100 SFRA has a valuable role.
References
1. CIGRE, “An International Survey on Failures in Large Power Transformers In Service.” (1983), Elec-tra NO 88, pp23-50.
2. W.H. Bartley, (1999), “An Analysis of Transformer Failures, Part 1” Locomotive, 73, 2, pp 4-7.
3. S.L. Nilssen and S. Lindgren, (1997), “ Review of Generator Step Up Transformer Failure Data”, EPRI Substation Conference, New Orleans.
4. A.L Rickley (1985) “Transformer Insulation Power Factors, A Progress Report” Minutes of the 52nd An-nual International Clients Conference, sec 6-201 5. J.A.Lapworth (1997) “CIGRE Working Group
12.18 Life Management of Transformers - An Activ-ity Overview.” ” Minutes of the 64th Annual Interna-tional Clients Conference, paper 8-8.
6. J.A.Lapworth and A.J. McGrail (1999) “Transformer Winding Movement Detection by Frequency Re-sponse Analysis” Minutes of the 66th Annual Inter-national Clients Conference, paper 8-14
agnostic Testing by Frequency Response Analysis”.
IEEE Trans PAS-97, No 6, pp 2144- 2153.
9. A.L.Rickley and R.E.Clark (1976), “Transformer Exciting Current Measured With Doble Equipment”
Minutes of the 43rd Annual International Clients Conference, sec 6-1101
10. M.F.Lachman, (1999) “Application of Equivalent Circuit Parameters to Off -line Diagnostics of Power Transformers”, Minutes of the 66th Annual Interna-tional Clients Conference, sec 8-10
Mr. Locarno received a BSEE from Northeastern University in Boston, MA in 1990. He worked as a startup engineer for the General Electric Co. power delivery systems. As a graduate of the GE fi eld engi-neering program he served in many roles; project manager for industrial applications resident engineer for IBM microchip division, and outage management for GE power generation services. Mr. Locarno has worked for Doble Engineering since 1996 and is currently a lead engineer in their new product technology group. Th e latest venture has been the develop-ment of a Swept Frequency Response Analyzer, for which he, (and others), hold Patent (pending review). Additionally, he acts as a project manager for their engineered strategies business unit which provides condition assessment and asset management to major utilities.